Google Inside

A few days ago, Google was granted a patent (#7,346,839) that will allow the association of documents based on historical data.

Here are the inventors noted for the patent,

• Acharya, Anurag (Campbell, CA, US)
• Cutts, Matt (Mountain View, CA, US)
• Dean, Jeffrey (Palo Alto, CA, US)
• Haahr, Paul (San Francisco, CA, US)
• Henzinger, Monika (Lausanne, CH)
• Hoelzle, Urs (Palo Alto, CA, US)
• Lawrence, Steve (Mountain View, CA, US)
• Pfleger, Karl (Mountain View, CA, US)
• Sercinoglu, Olcan (Mountain View, CA, US)
• Tong, Simon (Mountain View, CA, US)

Matt Cutts is on the team of inventors for this patent. Matt Cutts plays a large role of the Search Quality team for Google and also has a strong position with the Spam Abuse team at Google. Hopefully we will see this patent play out into how spam is combated within Google’s search. But next, let’s take a look at what the patent will actually do for Google.

What is claimed is:

1. A method, comprising: identifying a document; obtaining a plurality of types of history data associated with the document, the plurality of types of history data including at least two of: data relating to a document inception date, data relating to document changes, query analysis data, data relating to link-based criteria, data relating to anchor text, data relating to document traffic, data relating to user behavior, domain-related data, data relating to ranking history, user maintained or generated data, data relating to unique words, bigrams, or phrases in anchor text, data relating to linkage of independent peers, or data relating to document topics; and generating a score for the document based on the plurality of types of history data associated with the document.

The patent actually digs into how the algorithm will function, to a certain degree. Google filed for this patent late 2003 (12/31/2003), so the patent has been pending for a little over four years now. Seems slightly ironic to me that Google filed a patent around the idea of historic data and it has taken this long for the patent to actually be granted — almost as if the patent may render useless now, but I doubt that is the case.

Google’s search algorithm is continually undergoing changes to improve the search results, so it’s always looking for ways to improve on top of the improvements.